A reduction in residual energy of inter-frame prediction with motion compensation and improvement in encoding efficiency are achieved by using a region-dividing type adaptive interpolation filter that takes an edge property of a picture into consideration. An edge calculation unit calculates edge information from reference picture data designated by a motion vector. A region dividing unit divides an encoding target frame into a plurality of regions that are units to which interpolation filters are adaptively applied based on the edge information. A filter coefficient optimizing unit optimizes an interpolation filter for a fractional-accuracy pixel for each of the regions. A reference picture interpolating unit interpolates the fractional-accuracy pixel of a reference picture using the optimized interpolation filter, and a predictive encoding unit performs predictive encoding using motion compensation of fractional-accuracy.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A video encoding method using motion compensation of fractional-accuracy, the method comprising: a step of acquiring reference picture data designated by a motion vector and calculating edge information from the acquired reference picture data; a step of performing division of an encoding target frame into a plurality of regions that are units to which interpolation filters are adaptively applied, based on the edge information; a step of optimizing an interpolation filter for a fractional-accuracy pixel for each of the regions; and a step of interpolating the fractional-accuracy pixel of a reference picture using the optimized interpolation filter and performing predictive encoding using the motion compensation of fractional-accuracy, wherein the edge information is an edge angle extracted from the reference picture data designated by the motion vector and edge strength, and the division into the regions is performed depending on (a) whether or not the edge strength is higher than a threshold value and (b) if the edge strength is higher than the threshold value, a range of angles to which the edge angle belongs, the video encoding method further comprises a determining step of determining the threshold value based on a decoded picture obtained by performing decoding on an already encoded picture, and the determining step determines the threshold value in accordance with the number of the regions and an average value of previous edge strengths.
A video encoding method uses fractional-pixel motion compensation. It calculates edge information (angle and strength) from reference picture data designated by motion vectors. Based on this edge information, it divides the encoding target frame into regions. The regions are determined by whether the edge strength exceeds a dynamic threshold, and if so, the angle range to which the edge belongs. The method optimizes interpolation filters for fractional-accuracy pixels for each region. It then interpolates the reference picture using the optimized filters and performs predictive encoding. The threshold is determined based on a decoded, already-encoded picture, considering the number of regions and the average of previous edge strengths to adjust the filtering adaptively.
2. A non-transitory compute readable medium storing a video encoding program which causes a computer to execute the video encoding method according to claim 1 .
A non-transitory computer-readable medium stores a video encoding program. This program, when executed by a computer, performs a video encoding method that uses fractional-pixel motion compensation. It calculates edge information (angle and strength) from reference picture data designated by motion vectors. Based on this edge information, it divides the encoding target frame into regions. The regions are determined by whether the edge strength exceeds a dynamic threshold, and if so, the angle range to which the edge belongs. The method optimizes interpolation filters for fractional-accuracy pixels for each region. It then interpolates the reference picture using the optimized filters and performs predictive encoding. The threshold is determined based on a decoded, already-encoded picture, considering the number of regions and the average of previous edge strengths to adjust the filtering adaptively.
3. A video decoding method using motion compensation of fractional-accuracy, the method comprising: a step of decoding interpolation filter coefficients for a fractional-accuracy pixel; a step of calculating edge information from reference picture data obtained from a decoded motion vector; a step of determining a region to which decoding target data belongs among a plurality of regions which are units to which interpolation filters are adaptively applied, using the edge information; and a step of switching an interpolation filter for the fractional-accuracy pixel for each of the regions, interpolating the fractional-accuracy pixel of a reference picture, and performing predictive decoding using the motion compensation of fractional-accuracy, wherein the edge information is an edge angle extracted from the reference picture data designated by the motion vector and edge strength, and the regions are determined depending on (a) whether or not the edge strength is higher than a threshold value and (b) if the edge strength is higher than the threshold value, a range of angles to which the edge angle belongs, the video decoding method further comprises a determining step of determining the threshold value based on a decoded picture obtained by decoding an already encoded picture, and the determining step determines the threshold value in accordance with the number of the regions and an average value of previous edge strengths.
A video decoding method uses fractional-pixel motion compensation. It decodes interpolation filter coefficients for fractional-accuracy pixels. Edge information (angle and strength) is calculated from reference picture data obtained from a decoded motion vector. Based on this edge information, the method determines the region to which the decoding target data belongs. The regions are determined by whether the edge strength exceeds a dynamic threshold, and if so, the angle range to which the edge belongs. The appropriate interpolation filter for the fractional-accuracy pixel is selected based on the region. It then interpolates the reference picture and performs predictive decoding. The threshold is determined based on a decoded, already-encoded picture, considering the number of regions and the average of previous edge strengths to adjust the filtering adaptively.
4. A non-transitory computer readable medium storing a video decoding program which causes a computer to execute the video decoding method according to claim 3 .
A non-transitory computer-readable medium stores a video decoding program. This program, when executed by a computer, performs a video decoding method that uses fractional-pixel motion compensation. It decodes interpolation filter coefficients for fractional-accuracy pixels. Edge information (angle and strength) is calculated from reference picture data obtained from a decoded motion vector. Based on this edge information, the method determines the region to which the decoding target data belongs. The regions are determined by whether the edge strength exceeds a dynamic threshold, and if so, the angle range to which the edge belongs. The appropriate interpolation filter for the fractional-accuracy pixel is selected based on the region. It then interpolates the reference picture and performs predictive decoding. The threshold is determined based on a decoded, already-encoded picture, considering the number of regions and the average of previous edge strengths to adjust the filtering adaptively.
5. A video encoding apparatus using motion compensation of fractional-accuracy, the apparatus comprising: an edge calculation unit which acquires reference picture data designated by a motion vector and calculates edge information from the acquired reference picture data; a region dividing unit which performs division of an encoding target frame into a plurality of regions that are units to which interpolation filters are adaptively applied, based on the edge information; a filter coefficient optimizing unit which optimizes an interpolation filter for a fractional-accuracy pixel for each of the regions; a reference picture interpolating unit which interpolates the fractional-accuracy pixel of a reference picture using the optimized interpolation filter; and a predictive encoding unit which performs predictive encoding using the motion compensation of fractional-accuracy, wherein the edge information is an edge angle extracted from the reference picture data designated by the motion vector and edge strength, and the division into the regions is performed depending on (a) whether or not the edge strength is higher than a threshold value and (b) if the edge strength is higher than the threshold value, a range of angles to which the edge angle belongs, the video encoding apparatus further comprises a determining unit which determines the threshold value based on a decoded picture obtained by performing decoding on an already encoded picture, and the determining unit determines the threshold value in accordance with the number of the regions and an average value of previous edge strengths.
A video encoding apparatus uses fractional-pixel motion compensation. An edge calculation unit acquires reference picture data designated by a motion vector and calculates edge information (angle and strength). A region dividing unit divides the encoding target frame into regions based on the edge information. The regions are determined by whether the edge strength exceeds a dynamic threshold, and if so, the angle range to which the edge belongs. A filter coefficient optimizing unit optimizes interpolation filters for fractional-accuracy pixels for each region. A reference picture interpolating unit interpolates the reference picture using the optimized filters. A predictive encoding unit then performs predictive encoding. A determining unit determines the threshold value based on a decoded, already-encoded picture, considering the number of regions and the average of previous edge strengths.
6. A video decoding apparatus using motion compensation of fractional-accuracy, the apparatus comprising: an interpolation filter coefficient decoding unit which decodes interpolation filter coefficients for a fractional-accuracy pixel; an edge calculation unit which calculates edge information from reference picture data obtained from a decoded motion vector; a region determining unit which determines a region to which decoding target data belongs among a plurality of regions which are units to which interpolation filters are adaptively applied, using the edge information; a reference picture interpolating unit which switches an interpolation filter for the fractional-accuracy pixel for each of the regions and interpolates the fractional-accuracy pixel of a reference picture; and a predictive decoding unit which performs predictive decoding using the motion compensation of fractional-accuracy, wherein the edge information is an edge angle extracted from the reference picture data designated by the motion vector and edge strength, and the regions are determined depending on (a) whether or not the edge strength is higher than a threshold value and (b) if the edge strength is higher than the threshold value, a range of angles to which the edge angle belongs, the video decoding apparatus further comprises a determining unit which determines the threshold value based on a decoded picture obtained by decoding an already encoded picture, and the determining unit determines the threshold value in accordance with the number of the regions and an average value of previous edge strengths.
A video decoding apparatus uses fractional-pixel motion compensation. An interpolation filter coefficient decoding unit decodes interpolation filter coefficients. An edge calculation unit calculates edge information (angle and strength) from reference picture data obtained from a decoded motion vector. A region determining unit determines which region the decoding target data belongs to based on the edge information. The regions are determined by whether the edge strength exceeds a dynamic threshold, and if so, the angle range to which the edge belongs. A reference picture interpolating unit selects the appropriate interpolation filter for the fractional-accuracy pixel based on the region and interpolates the reference picture. A predictive decoding unit performs predictive decoding. A determining unit determines the threshold value based on a decoded, already-encoded picture, considering the number of regions and the average of previous edge strengths.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
August 10, 2011
March 28, 2017
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.